Isotopic sediment ˜ngerprinting to correlate land use changes and … · 2014. 6. 13. · *...

Overview of projectOverarching project goals:

* Determine relationships among land use, sediment yield, and type of erosion for gauged watersheds in Yunnan

* Determine the e�ects of grain size and mineralogy on fallout radionuclide retention in sediments

* Determine the di�erences between monsoon and dry season sediments to better understand erosion type and patterns in Yunnan

Project goals:* Characterize background erosion rates using 10Be in �uvial sediment* Understand erosion rates in the context of modern land use* Understand erosion rates in the context of modern sediment yield

Methods:* In situ 10Be analysis of 54 samples collected in summer 2013 (Fig. 1)* Network analysis tracking sediment as it moves downstream* Correlation analysis of erosion rates and geomorphic parameters

Preliminary results:* Erosion rates range from 12 ± 1 to 195 ± 18 m/My* Slope is the strongest regressor with all samples (r2 = 0.37, p < 0.001)* No correlation with size, elevation, mean annual precipitation* Erosion rates for individual watersheds correlate less well with slope* Network analysis of samples upstream and downstream of con�uences

suggest using 10Be data to estimate sediment yield is inaccurate (44% under to 92% over actual yields)

Project goals:* Characterize land use for 3 basins sampled in 2013 (Fig. 1) into forested,

cultivated, and bare earth* De�ne watersheds for samples collected in 2014* Characterize land use for samples collected in 2014 into forested,

cultivated, and bare earth

Methods:Land use classi�cation steps* Acquire clear satellite imagery* Calculate radiances and correct for atmospheric e�ects* Use the ENVI Feature Extraction module for land use classi�cation* Check land use

classi�cation against global land use datasets (such as GLOBcover) and �eld photos to assess accuracy

Drainage basin extraction* Acquire high resolution DEMs (ASTER GDEM or HydroSHEDS)* Correct DEMs for errors (including sinks and internal drainages)* Calculate �ow direction and determine basin outlets* Calculate drainage basins using ArcGIS watershed tool

Project goals:* Compare modern sediment yield with long-term erosion rates at 26

Chinese hydrology stations (Fig. 5)* Use 4-isotope system to �ngerprint depth of erosion (Fig. 13)* Correlate depth of erosion to upstream land use

Methods* Collected samples in January 2014* Samples are being analyzed for 137Cs, 210Pb, meteoric 10Be, and in situ 10Be

during summer and fall 2014* Analysis of results during fall 2014 and spring 2015

Project goals: * Determine the e�ects of drainage tiles on depth of erosion in the

Vermilion watershed, northern Ohio

Methods:* Analyze samples collected by Jenny Bower (OC ‘13) for 137Cs (Fig. 10)* Model drainage tile locations using soil type and land cover (completed

by Jenny Bower (OC ’13) (Fig. 11)* Determine sample mineralogy using XRD (Fig. 12)

Preliminary results:* 137Cs activity is correlated with

drainage tile density, suggesting that drainage tiles promote shallower (surface) erosion.

Project goal:* Determine the e�ects of grain

size and mineralogy on retention of fallout radionuclides in sediment

Methods:* Measure meteoric 10Be, 137Cs, and 210Pb in 5

grain size fractions from 6 in channel sites sampled during the 2014 �eld season (Fig. 8)

* Natural delivery experiment to see how 5 di�erent sheetsilicates, 5 grain sizes of quartz, and 5 grain sizes of natural river sediment accumulate fallout radionuclides (7Be and 210Pb) (Fig. 9)

Project goals:* Use isotopes and

mineralogy to determine di�erences in sediment source during the wet and dry season

Methods:* At all sample sites we collected in

channel (dry season) and overbank (wet season) deposits

* 137Cs and 210Pb analysis of paired (in channel and overbank) samples

* XRD analysis of paired samples

Preliminary results:* Dry and wet season sediment

sources often di�er (Fig. 7)

Project goals:* To determine how short-lived radionuclide activity changes throughout

one watershed (basin 49, Fig. 1)* To connect the network analysis of 137Cs and 210Pb activity to qualitative

changes in land use

Methods:* 137Cs and 210Pb analysis of detrital sediments* Qualitative analysis of land use from �eld photos and Google Earth* Network analysis of 137Cs and 210Pb activity

Preliminary results:* Central tributaries have the

lowest activity* Low activity may be

related to land use* Mainstem sediment appears to be sourced mainly from eastern and western tributaries

Project goals:* Connect patterns in sediment yield to land use changes and policies* Explore patterns in short-lived radionuclide activity in sediments and

historical and current land use for one small watershed (basin 35, Fig. 1)

Methods:* 137Cs and 210Pb analysis of detrital sediments* Land use/land cover analysis from historic satellite images* Analysis of Chinese sediment yield and discharge data

Preliminary results:* 137Cs is only found in two high,

mountainous watersheds which have been pasture for the entire study period

* 210Pb activity does not correlate with land use, slope, or elevation

* Two peaks in sediment yield do not correlate with rainfall and may be related to land use policies (Fig. 2)

Isotopic sediment �ngerprinting to correlate land use changes and erosionFaculty: Amanda Schmidt, Oberlin College in collaboration with Paul Bierman (University of Vermont) and Dylan Rood (Scottish Universities Environmental Research Centre)

Students: Oberlin College - Dominic Fiallo, Gabriela Garcia, Joseph Martin, Yue Qiu, Adrian Singleton, Sylvia WoodmanseeUniversity of Vermont - Thomas Neilson, Veronica Sosa Gonzalez

China

Myanmar

Hanoi

XiaguanKunming

Chengdu

Calcutta

35

49

11

105°0'0"E100°0'0"E95°0'0"E90°0'0"E

32°30'0"N

30°0'0"N

27°30'0"N

25°0'0"N

22°30'0"N

120°0'0"E100°0'0"E80°0'0"E

45°0'0"N

30°0'0"N

15°0'0"N

0°0'0"

Figure 1: We sampled at 54 sites in 3 watersheds (in orange) in May 2013. Watersheds are all tributaries to the Mekong.

Yue Qiu - Sediment yield, 137Cs, 210Pb, and land use

Figure 2: Sediment yield generally tracks with rainfall except for high sediment yield in 1979 and 1984.

Dominic Fiallo and Gabriela Garcia - Network analysis of 137Cs and 210Pb in one watershed

0 20 40 60 80 100−4

−2

0

2

4

6

8

10

12x 10−3

Normalized area

210 P

b A

ctiv

ity [B

q/g]

Mainstem

Overbank

Center upstream

East tribs

West tribs

Central tributaries

Mainstem

Westerntributaries

Easterntributaries

Figure 3: The central tributaries have statistically signi�cantly lower 210Pb activity than the eastern or western tributaries or the mainstem. This may be due to intensive farming, including tobacco plantations, in the central tributaries.

Joseph Martin - GIS and remote sensing

Basin 35

Basin 11

Basin 49

Figure 4: These maps show the panchromatic band of the Landsat 8 image being used for land use classi�cation. See �gure 1 for basin locations in Yunnan.

Veronica Sosa Gonzalez - Sediment yield, erosion depth, and long-term erosion rates

Figure 5: (a) Samples (stars); (b) High water labels at a hydrology station; (c) A hydrology station; (d) Sieving sediments at a sample site.

^^

^

^

^̂^

^

^^^

^

^^

^

^

^̂^

^

^^ ^

^

^

^

China

Myanmar (Burma) VietnamLaos

a

dc

b

Dominic Fiallo - Monsoon vs dry season sediment sources

Adrian Singleton -

Grain size and mineralogy

e�ects on nuclide retention

Figure 6: Dom with the Oberlin XRD during training to use the machine

Figure 7: Basins 49 and 15 show deeper sediment sourcing during monsoons compared with dry seasons. Basin 35 has no change and basin 11 has shallower sources.

11 35 49 15

Pb

activ

ity [B

q/g]

Overbank (wet season)

In channel (dry season)

Higher overbank activity suggests shallower wet season sediment source

Lower overbank activity suggests deeper wet season sediment source

Wet and dry season sources likely similar

Figure 8: (a) Sampling for grain size experiment in the �eld in January 2014. (b) A grain size sample site.

Figure 9: Natural delivery experiment. Samples are exposed for one month then brought inside for counting. Rainwater is also collected and measured.

Sylvia Woodmansee - Drainage tiles and erosion

Figure 10: Sylvia with our germanium detector, Harbin. Harbin is named because the detector has to stay cold with LN2 and Harbin is a very cold city in China where Amanda used to live.

Drainage Tile Distribution in the Vermilion River Watershed

0 7.53.75 Miles Ü

LegendRiver

Drainage Tiles

2 miles

Figure 11: Modeled drainage tile locations (brown) in the study area. Drainage tiles are modeled as places with agricultural land use and soils classi�ed as “prime farmland if drained”. Note: Although Jenny has graduated, she is involved

in this project and it is in collaboration with her.

Overview

Figure 13: Using a combination of the four isotopes shown in this �gure allows us to �ngerprint type

and depth of erosion better than using any single system. We will correlate the nuclide

�ngerprint of each sample to the upstream land use to determine

land use e�ects on type of erosion.

Sheetwash/ Shallow Rills

Rills, Shallow Gullies, Shallow

Landsliding, Bank Failure/

Channel Migration

Deep Gullies/ Landslides

Dep

th (m

)1

2

3

Increasing Concentration Dominant type of erosion

in situ 10Be

meteoric 10Be“bulge” profile

137Cs210Pb

meteoric 10Beexponential profile

Figure 14: This project takes advantage of an unusually complete record of sediment yield and discharge for 26 gauging stations in Yunnan (such as the one shown) with at least 5 years of daily discharge and sediment concentration data. Field work was completed in 2013 and 2014.

b

a

Coun

ts/s

ec

Degrees

Figure 12: Preliminary XRD data for one sample. We expect areas with more tiles to have more weathered material.

Thomas Neilson - long term erosion rates